The goal of this work is the development of a practical, non-penetrating somatosensory neural interface at the level of the dorsal root ganglia (DRG), for use as sensory feedback. The DRG is an ideal location to record somatosensory neural signals which convey body-state information such as tactile and proprioceptive feedback from the limbs. These signals can be used as control signals in closed loop functional electrical stimulation (FES) applications in which the position of a limb is decoded and used to regulate the FES system to adapt to fatigue or the addition of a load or to perform complex, multi-joint movements smoothly. Internal neural interface sensors, such as these may be minimally obtrusive and integrate easily with implanted neuroprostheses while not requiring the large number of sensors or regular donning and calibration that external sensors do. Our lab has shown the ability to decode limb position with high accuracy from recordings of primary afferent neural activity with penetrating microelectrodes inserted in lumbar DRG, and has applied these signals as feedback within closed-loop FES control of the lower limb. However, the efficacy of long-term recordings from these penetrating electrodes has yet to be established, and recordings with penetrating electrodes in humans are challenging to obtain. We hypothesize that recordings from the surface of the DRG may be sufficient for extracting detailed information about limb position and may provide a more practical route for clinical evaluations. Unlike other neural structures, cell bodies are packed closely under the DRG perineum, making it an ideal candidate for obtaining activity from individual cells without using penetrating electrodes. Also, this surface approach may have a higher efficacy in long-term recordings, due to a reduced tissue response, than penetrating electrodes. Recently we showed that recordings from the surface of the DRG, with non-penetrating electrodes, can yield neural signals that can be used to predict the state of the limb, in animal studies. We will evaluate two specific aims in the research planned in this study. Aim 1: In animal experiments, we will evaluate the ability of neural recordings from non-penetrating electrodes on the L6 and L7 DRG to provide functional closed loop control of FES controlled lower limb stepping movements. This study will establish the utility of this approach, by indicating whether a sufficient range of neural activity can be recorded from, to obtain feedback towards functional limb movements. Aim 2: In intraoperative human experiments, we will evaluate the ability of non-penetrating electrodes placed on the surface of exposed DRG to record neural activity that predicts mechanical and vibratory stimulation applied to the lower leg. This study will obtain the first recordings from DRG in humans and demonstrate that we are able to decode human DRG activity. Success in this proposed work will be an important advance towards development of closed-loop neuroprostheses based on this non-penetrating DRG approach and will drive future studies on extended duration animal and human recordings using optimized electrodes.